Value of “true” leukocytosis formula in determining total knee arthroplasty prosthetic joint infection in the community hospital setting

Abstract

BACKGROUND

With the advent of recent point of care testing to evaluate for prosthetic joint infection (PJI) such as alpha-defensin and calprotectin, the diagnosis of PJI has reached a new paramount. The most widely accepted diagnostic criteria in the literature at this time is the international consensus meeting (ICM) score by Parvizi et al in 2018. One of the proposed limitations of the ICM criteria is determining if fluid analysis can be trusted for bloody joint aspirations. A proposed formula to correct red blood cells in aspirations was described in 2008 but has not been explored since then.

AIM

To determine if the “true leukocytosis” formula for prosthetic joint aspirations is useful in determining PJIs.

METHODS

Of 158 cases with International Classification of Diseases, 10^th revision diagnoses of total knee arthroplasty PJIs from 2020-2021 from a single health system were reviewed and preoperative ICM scores were calculated. Inclusion criteria involved preoperative scores indicating infection, first time diagnoses, and culture positive infections. Culture positivity remained our gold standard for infection confirmation. We compared sensitivity, specificity, positive predictive value, negative predictive value, and area under the curve for preoperative ICM score and corrected preoperative ICM score.

RESULTS

Of 81 cases met the inclusion criteria and had preoperative ICM scores calculated. Of the 81 cases that met inclusion criteria, no cases were found to have different preoperative ICM scores after applying the correction formula for “true leukocytosis” proposed by Ghanem et al in 2008.

CONCLUSION

Given the lack of variance in ICM score following the proposed correction formula, it appears that the proposed “true leukocytosis” formula is not a useful adjuvant in treatment decision making.

Key Words: Periprosthetic joint infection; International consensus meeting criteria; Criteria; Total nucleated cell count; True leukocytosis; Total knee arthroplasty

Core Tip: In this study we examine the usefulness of applying a correction formula for synovial fluid aspirations that contain red blood cells when determining prosthetic joint infections (PJI) using the international consensus meeting criteria. We compare the international consensus meeting criteria for PJI score both prior to and after applying the correction formula to evaluate if there would be any change to diagnosis in patients with concern for PJI.

INTRODUCTION

Prosthetic joint infection (PJI) is a devastating complication following total knee arthroplasty (TKA). While the overall rate of PJI in TKA appear to have remained consistent over the past three decades[1], the incidence of primary TKA has continued to rise and is projected to increase to a total of 3.48 million procedures a year in the United States by 2030 with other estimates suggesting that the incidence of TKA could be even higher by 2040[2-5]. With this increase in primary TKA there will also come an increase in incidence of revision TKA which will place a large burden on the healthcare system from both a clinical and economic standpoint. Over a 10-year period, Livshetz et al[3] projected that the cost of revision total knee will create up to 4.1 billion United States dollars in hospital charges[6]. Given that infection is the most common cause for revision TKA, accounting for up to 25% of revisions, we can therefore also expect an increase in PJI with the expected increase in revisions[7].

While this complication is well described in the literature there remain several identifiable problems in both the diagnoses and treatment of PJI. Many criteria have been suggested and subsequently revised for the diagnoses of PJI: Musculoskeletal Infection Society criteria (2011), international consensus meeting (ICM) criteria[8] and European Bone and Joint Infection Society criteria (2021)[8-11]. The criteria with the highest sensitivity, specificity, and internal validation are the 2018 criteria proposed by the ICM[8]. Certain issues remain despite the clear guidelines listed in the ICM criteria, specifically in the community setting[12].

The ICM criteria currently utilizes major and minor criteria for diagnoses. The major criteria being a draining sinus track that probes down to prosthesis or a positive culture on aspiration. If either of the major criteria are met, the prosthetic joint is assumed to be infected. The minor criteria requires that the affected joint be aspirated, and the fluid sent for analysis of total nucleated cell count (> 3000 = 3 points), neutrophil percentage (> 80% = 2 points) , presence of alpha-defensin (3-points), leukocyte esterase (3-points), and C-reactive protein (CRP, 3-points); it also calls for serum erythrocyte sedimentation rate (> 30 mm/hour = 1 point) and serum CRP (> 1.0 mg/dL = 2 points) be acquired. Each of these minor criteria are weighted and a cumulative preoperative weighted score of > 6 is assumed infection for the preoperative criteria[8]. In the community setting, clinicians are limited by what their labs can test for. As a result, many PJIs at these community hospitals fall into the indeterminant range, 2-5 points, and as a result this may delay or even prevent appropriate diagnosis and treatment.

While these three diagnostic modalities have been widely utilized over the past decade with support from the musculoskeletal infection society, several other proposed diagnostic tests and criteria exist. One of these proposed criteria is to solely rely on the intra-articular enzyme alpha-defense and intra-articular CRP[13]. This notion, originally proposed by Deirmengian et al[14] has shown good sensitivity and specificity in diagnosing PJI when compared to gold standard of ICM criteria[15-18]. However, to avoid confusion, the proposed CRP-Deirmengian criteria has been excluded from this study, and only the ICM diagnostic criteria was utilized as it is the current gold standard.

In the community setting the clarity and reliability of the cell counts and neutrophil percentages on aspirate are of vital importance. Unfortunately, another major limitation of aspiration in many cases is the presence of red blood cells (RBCs) which has been suggested to influence the fidelity of the total nucleated cell count. While it has been widely implied that the presence of RBCs in joint aspirate can inflate the total nucleated cells (TNC), there has been limited research into the actual effects RBCs have on cell count. Deirmengian et al[14] released a study in 2020 that suggested samples diluted with large amounts of blood or saline should not be utilized in defining infection. This study further suggests that without fidelity of the cell counts on aspiration, many patients who end up in the “indeterminate” group when applying the ICM criteria, will have delayed diagnosis or incorrect treatments. In this study the term “grossly bloody” was defined as any fluid sample that appeared opaque and sanguinous on physical evaluation in the syringe.

The major issue is that without fidelity of synovial fluid analysis, it is impossible to meet the minimum criteria for PJI per the ICM criteria[8], which lead our cohort to investigate potential means to adjust cell counts in the setting of “bloody aspirations”. In 2008, Ghanem et al[19] published a possible correction formula to identify the “true leukocytosis” present in the prosthetic joint. This formula offers a new potential for validating the fidelity of total nucleated cell count on aspiration and since the time of publication in 2008, there has been no follow-up research applied to this proposed correction formula. Additionally, to our knowledge there has not been a definitive stance on whether this correction formula is useful in the clinical setting. The purpose of the present study is to determine the true clinical utility of the “true leukocytosis formula” through comparison with the ICM confirmed infections. By doing so, we aim to provide insight for clinicians in the community setting with limited access to various diagnostic testing modalities regarding the usefulness of this compensatory formula in the setting of a “bloody aspiration”.

MATERIALS AND METHODS

A retrospective chart review was performed from January 2020 to December 2021 to evaluate patients with culture positive periprosthetic joint infection after a TKA with confirmed PJI by ICM score and culture positivity. Patient demographics including age, body mass index, gender, diabetes, hypertension, hyperlipidemia, obesity, atrial fibrillation, obstructive sleep apnea, chronic kidney disease anxiety and depression were recorded. Preoperative ICM score was calculated for each patient without alpha-defensin or intraarticular CRP, as these tests are not available at our institution[8]. Only patients with preoperative ICM scores indicating infection were included in the experimental group. We then calculated the corrected ICM score by using the cell count in conjunction with the equation from Ghanem et al[19] to identify the “true leukocytosis” present in the prosthetic joint. The proposed equation was: White blood cell (WBC)_adjusted = WBC_fluid – [WBC_blood × (RBC_fluid/RBC_blood)]; where WBC stands for white blood cells and RBC stands for red blood cells. We used culture positivity as our gold standard for proven infection. We compared sensitivity, specificity, positive predictive value, negative predictive value, and area under the curve for preoperative ICM score and corrected preoperative ICM score. Culture results were also reported based on bacterial involvement.

Patients who experienced a first time PJI who underwent operative intervention with debridement, antibiotics and implant retention, or explant with placement of antibiotic spacers were included. Readmission and reinfection were not evaluated as this was deemed out of scope. Patients were excluded if recurrent infection was found, or if cultures were negative. For the purpose of this study the term “grossly bloody” refers to any synovial fluid sample that appeared opaque and sanguinous on physical evaluation in the syringe.

Descriptive statistics included reporting the numerical variables as mean, standard deviation and range, and the categorical variables as number and percent. Positive predictive value or negative predictive value were not reported due to the 100% matching between the comparison variables. All the analysis were performed with the use of SAS software, version 9.4 (SAS Institute, Cary, NC, United States).

RESULTS

Of the 158 cases, 81 were included in the study. Of these 81 cases, zero cases were found to have a change in the preoperative ICM score following application of the “true leukocytosis” formula proposed by Ghanem et al[19]. The results met clinical significance as 100% of patients remained at their preoperative ICM scores. There were no secondary outcomes measured for this study. The results of the patient demographics, criteria derived from the ICM criteria, results from our patients fluid aspirations, and the species of bacteria isolated from the aspirations can all be found in Tables 1, 2, 3 and 4.

Table 1 Preoperative international consensus meeting criteria scoring system.

	Cut off	Score1
Serum CRP	> 1 mg/dL	2
Serum D-dimer	> 860 ng/mL	2
Serum ESR	> 30 mm/hour	1
WBC (synovial)	> 3000 cells/μL	3
Alpha-defensin (synovial)	Qualitative	3
Leukocyte esterase (synovial)	Qualitative	3
Neutrophil (synovial, %)	> 80%	2
CRP (synovial)	> 6.9 mg/L	1

¹Adapted from Parvizi et al[11].

CRP: C-reactive protein; ESR: Erythrocyte sedimentation rate; WBC: White blood cell.

Table 2 Patient demographics (n = 81), n (%).

Characteristic	Value
Age (mean ± SD, range)	68.3 ± 9.1, 52-88
Gender
Male	46 (56.79)
Female	35 (43.21)
BMI (mean ± SD, range)	35.3 ± 8.5, 19.2-61.9
10-year survival (mean ± SD, range, %)	65.4 ± 17.0, 0-95.9
Comorbid conditions
DM	21 (25.93)
HLD	34 (41.98)
HTN	64 (79.01)
Obesity (BMI > 35)	38 (46.91)
AFIB	9 (11.11)
CKD	12 (14.81)
Smoking	3 (3.70)
Anxiety/depression	40 (49.38)
Sleep apnea	25 (30.86)

BMI: Body mass index; DM: Diabetes mellitus; HLD: Hyperlipidemia; HTN: Hypertension; AFIB: Atrial fibrillation; CKD: Chronic kidney disease.

Table 3 Average results from fluid analysis.

	Average	Range
RBC serum cells (× 106)	4.02	2.1-5.1
RBC synovial	183761	4843-4063600
WBC serum cells (× 103)	10.9	6.3-31.7
TNC synovial	66076.5	117-494920
Polymorphonuclear cells (%)	87.37407407	74-100
C-reactive protein serum	16	0.6-33.9
Erythrocyte sedimentation rate (mm/hour)	60.5	6.8-140

RBC: Red blood cell; WBC: White blood cell; TNC: Total nucleated cell.

Table 4 Species of bacteria isolated on intra-operative cultures.

Bacterium isolated	Number of patients	Percent of patients
MRSA	14	17.3
Methicillin sensitive staphylococcus aureus	21	25.9
Coagulase negative Staphylococcu spp.	7	8.6
Enterobacter spp.	4	4.9
Pseudomonas spp.	3	3.7
Proteus spp.	1	1.2
Corynebacterium spp.	3	3.7
Enterococcus Faecelis	8	9.9
Bacteroides spp.	1	1.2
Group A streptococcus spp.	5	6.2
Peptostreptococcus spp.	1	1.2
Group B hemolytic streptocossu	5	6.2
Escherichia coli	2	2.5
Mycobacterium spp.	1	1.2
Propionibacterium acnes	1	1.2
Serratia spp.	1	1.2
Bacillus spp.	3	3.7

MRSA: Methicillin-resistant staphylococcus aureus.

DISCUSSION

TKA has seen increasing rates of utilization leading to increases in both revision TKA and PJI. Furthermore, as the ability to perform TKA becomes increasingly accessible in the community setting, many surgeons with limited resource access will be tasked with diagnosing PJI accurately. In the hospital the present study is based in, alpha defensin, synovial CRP, and leukocyte esterase are not readily available tests after an aspiration. In the setting of a bloody aspiration, without the power of these confirmatory tests, the diagnostic and therefore treatment decisions may be questioned. In this study we looked at determining whether the Ghanem et al[19] proposed correction equation for bloody aspirations of TKAs was useful in the diagnosis of PJI for TKA when a bloody tap is encountered. It was our hope that having this additional data point would be helpful in maintaining confidence in the validity of our diagnosis in a community setting where adjuvant testing is not readily available.

The patients in the study had an ICM score calculated from preoperative fluid analysis and lab values. The intra-articular WBC was then subjected to Ghanem et al[19] proposed “true” leukocytosis equation which showed that zero patients had a change to their ICM score following application of the “true” leukocytosis equation. Unfortunately, our team has interpreted this to mean that this proposed equation is not a useful adjuvant in routinely diagnosing PJI. While our cohort was relatively small in this study, we believe the equation’s proposed formula would likely require increasingly large amounts of RBCs (> 5000000/mL) in joint fluid aspirate to create a substantial change to the intra-articular WBC count. This, however, does appear to coincide with Deirmengian et al[14] study from 2020 that suggested nucleated cell counts began to lose their fidelity in samples with greater than 1000000 RBCs/mL. In this context, perhaps aspirations which are frankly bloody and present with greater than 1000000 RBC would benefit from the correction.

The results of this study suggest that the previously hypothesized effects of RBC in aspirate on cell count may not be as clinically significant in samples with < 1000000 RBCs/mL. Given that there was minimal change in TNC before and after correction using the proposed equation, in our opinion, reflects that more confidence can be placed in TNC following aspirations even if they are grossly bloody. At this time, it appears that this equation is not as valuable as originally thought for routine aspirations with moderate amounts of RBCs. Given the increasing availability of newer diagnostic testing, such as intra-articular alpha-defensin, the absolute necessity of traditional fluid analysis may become less favorable when compared to these newer tests. If that is the case, this may be an area of study that while interesting may yield diminishing returns. However, until these newer tests are available in all settings, aspirate fluid analysis and utilization of the ICM score will continue to be the gold standard.

CONCLUSION

We conclude that there appears to be little to no effect on ICM criteria scores when the proposed correction formula applies. When we compared sensitivity and specificity of corrected score to the gold standard of culture positivity, there was no change in results compared to the uncorrected scores[9]. This could point to the fact that RBCs in aspirations present little to no effect on the cell count, contrary to previous thoughts[20]. For this reason, we propose that no correction to cell count needs to be done prior to diagnosis of PJI for TKA in which the sample is not grossly bloody.